BACKGROUND: Neck dissection is a common procedure for treatment of head and neck cancer. Removal or damage to the spinal accessory nerve can lead to "shoulder syndrome", characterized by shoulder droop, winged scapula, weak abduction, inability to shrug and a dull ache with pain localized to the shoulder. Though modified surgical procedures have been developed which spare the spinal accessory nerve, these procedures do not prevent shoulder syndrome entirely. There are no effective treatments for shoulder syndrome and many patients live with significant long-term pain and disability. Acupuncture is widely used for pain and there is evidence of its effectiveness for both acute and chronic pain disorders. These considerations warrant further research on acupuncture for shoulder syndrome following neck dissection. OBJECTIVES: 1. To determine whether a Phase III trial of acupuncture for shoulder pain and dysfunction in cancer patients treated by neck dissection is warranted; 2. To determine whether such a trial is feasible in terms of accrual rate, drop-out rate, use of off-study treatment and compliance; 3. To determine the most sensitive outcome measure; 4. To provide data necessary (or sample size calculation. METHODS: Head and neck cancer patients complaining of shoulder pain and dysfunction at least three months after lack dissection will complete two visual analog scales of shoulder pain on movement, one week apart. They will also be assessed by a physical therapist using the Constant-Murley score, a validated shoulder pain and dysfunction assessment instrument. Patients will be randomized to eight biweekly treatments of acupuncture or to no additional care. Outcome will be assessed two weeks after the end of treatment. Change in pain and Constant-Murley will be compared between groups by analysis of co-variance. At the end of the trial we will have data to address each of our aims as follows: Is a Phase III indicated? Data: Confidence interval for the difference between group means. Aim: Is a Phase III trial feasible? Data: Accrual rate; drop-out rate; compliance with treatment; use of off-study treatments. Should Constant-Murley or VAS be used? Data: Effect sizes calculated by adjusted mean difference divided by adjusted standard deviation Aim: Sample size calculation for Phase III. Data: Standard deviations, mean of control group, correlation between and within baseline and follow-up measures.